Mon. Not. R. Astron. Soc. 000, 000–000 (0000) Printed 8 October 2018 (MN LATEX style file v2.2) On the 3 dimensional structure of edge-on disk galaxies Michael Pohlen1,2 ⋆, Saleem Zaroubi1, Reynier F. Peletier1, & Ralf-Jurgen¨ Dettmar2 1 Kapteyn Astronomical Institute, University of Groningen, PO Box 800, NL-9700 AV Groningen, The Netherlands 2 Astronomisches Institut, Ruhr-Universit¨at Bochum, Universit¨atsstrasse 150, D-44780 Bochum, Germany 8 October 2018 ABSTRACT A simple algorithm is employed to deproject the two dimensional images of a pilot sample of 12 high-quality images of edge-on disk galaxies and to study their intrinsic 3 dimensional stellar distribution. We examine the radial profiles of the stars as a function of height above the plane and report a general trend within our sample of an increasing radial scalelength with height outside of the dustlane. This could be explained by the widespread presence of a thick disk component in these galaxies. In addition, the 3-dimensional view allows the study of the vertical distribution of the outer disk, beyond the break region, where we detect a significant increase in scalelength with vertical distance from the major axis for the truncated disks. This could be regarded as a weakening of the “truncation” with increasing distance from the plane. Furthermore, we conclude that the recently revised classification of the radial surface brightness profiles found for face-on galaxies is indeed independent of geometry. In particular, we find at least one example of each of the three main profile classes as defined in complete samples of intermediate to face-on galaxies: not-truncated, truncated and antitruncated. The position and surface brightness that mark the break location in the radial light distribution are found to be consistent with those of face-on galaxies. Key words: galaxies: photometry – galaxies: structure – galaxies: fundamental parameters – galaxies: evolution – galaxies: formation galaxies: individual: ESO 380-019, ESO 404-018, FGC2339, IC4871, NGC522, NGC1596, NGC3390, NGC4179, NGC5290, NGC5290, NGC5981, UGC 10459 arXiv:astro-ph/0703768v1 29 Mar 2007 1 INTRODUCTION (e.g. Patterson 1940; de Vaucouleurs 1959) noticed that the radial light distribution of stellar disks is well described by a simple expo- To study galaxy formation and evolution it is often helpful to imag- nential, albeit the lack of a clear physical explanation for this fact. ine that galaxies are made up of a small number of separate com- Later on, Freeman (1970) called these disks Type I – as opposed to ponents. Surface photometry of galaxies provides a very simple some other galaxies called Type II – showing a break in their radial way of parameterising these various components. By modelling the profile with a downbending steeper outer region, or equivalently, a surface-brightness distribution of a galaxy and thereby parametris- flatter inner region. ing the individual components, we obtain a common ground to measure, compare, and sort large samples of galaxies. Using the edge-on view, van der Kruit (1979) discovered Owing to their flat geometry the random orientation of a spe- that the exponential decline does not continue to the last mea- cific galaxy disk provides us with different projected views of sured point, but is truncated after several radial scalelengths. Af- the, presumably, common underlying 3 dimensional distribution. In terwards, with the advent of CCD technology, several groups particular, one gets complementary views from observing face-on (Barteldrees & Dettmar 1994; Pohlen et al. 2000b; Florido et al. or edge-on galaxies. In the face-on view, one studies the radial dis- 2001; de Grijs et al. 2001; Pohlen 2001; Kregel et al. 2002) stud- tribution of the starlight discussing, e.g., non-axisymmetric features ied larger samples of edge-on galaxies to find out if and at what such as bars, rings, or spiral arms. In the edge-on view on the other radius this truncation occurs. hand, one is able to study the vertical distribution – e.g., disentangle More recently, Erwin et al. (2005), Erwin et al. (2006), and thick disk, thin disk and stellar halos. Due to a longer line-of-sight Pohlen & Trujillo (2006) have analysed surface brightness profiles projection, the edge-on view allows in principle the exploration of of face-on to intermediate inclined galaxies and revised our view the radial profile to larger distances. on disk truncations. They identified three basic classes of surface Early on, from face-on galaxy observations, several authors brightness profiles depending on the apparent break features or lack of them: 1) The well known Type I that has an exponential profile, with no break. 2) Type II with a ‘downbending break’. 3) A com- ⋆ E-mail: [email protected] pletely new class, called Type III, also described by a broken ex- c 0000 RAS 2 M. Pohlen, S. Zaroubi, R.F. Peletier, & R. -J. Dettmar ponential but with an upbending profile beyond the break. In this with the line-of-sight integration (see e.g. Pohlen et al. 2004a) we view the classical truncations described initially in the pioneering decided to deproject a sample of edge-on galaxies. This can only work of van der Kruit (1979) and van der Kruit & Searle (1981) are be done using some geometric assumption. Here we choose the as- in fact part of a generalised Freeman Type II class (Freeman 1970). sumption of axial symmetry. This will allow us to gain insight into In this picture truncated galaxies show a surface brightness profile the three-dimensional galactic stellar disk structure by bringing to- with a shallow inner and a steeper outer exponential region sepa- gether the classical face-on and edge-on views. rated at a relatively well defined break radius. The vertical structure may well be the key to understand the These results are in agreement with those reported by nature of the different breaks in the surface brightness profiles, Pohlen et al. (2002a), who obtained deep photometry of three face- by setting additional constraints on the various physical explana- on galaxies and clearly demonstrated that the sharp cut-offs1 fit- tions put forward. For example, some recent star formation mod- ted to profiles of edge-on galaxies (e.g. by van der Kruit & Searle els (Elmegreen & Hunter 2006; Li et al. 2006) are now able to pro- 1981; Barteldrees & Dettmar 1994; Pohlen et al. 2000b) are in fact duce the observed broken exponential structure, but, alternatively, not complete but better described by a broken exponential. Even be- Debattista et al. (2006) also find downbending breaks using purely fore, de Grijs et al. (2001) and Pohlen (2001) noted that for edge- collisionless N-body simulations. Once the nature of the galactic on galaxies the surface brightness distribution does not disappear stellar disk is determined it will be fully justified to utilise these asymptotically (‘vertically’) into the background noise. Interest- breaks for comparison of galaxies at various redshifts as done by ingly, Hunter & Elmegreen (2006) find a similar broken exponen- P´erez (2004), Trujillo & Pohlen (2005) or Tamm & Tenjes (2006), tial shape for many objects in their large sample of irregular galax- and explain their appearance in cosmological simulations as shown ies. by Governato et al. (2007). The redefinition of the truncation as a point at which the two The remainder of this paper is organised as follows. Section 2 exponential fits meet solves, among other things, the apparent con- describes our sample selection and data. In Section 3 we introduce tradiction between simulations by Holley-Bockelmann & Mihos our method to deproject the edge-on galaxies. The results are given (2001) and the edge-on observations of Barteldrees & Dettmar in Section 4, we discuss them in Section 5 and conclude in Sec- (1994) and Pohlen et al. (2000a). The latter authors find that ‘trun- tion 6. In Appendix A we give detailed comments for all our galax- cations’ are as ‘close’ as twice the radial scalelength from the centre ies and show in Appendix B radial surface brightness profiles and in contradiction with the simulations. Furthermore, ‘early’ trunca- isophote maps for all galaxies. tions (i.e., at R < 2h) of edge-on galaxies have been always ques- tioned by the ‘face-on∼ viewers’, who had never observed a face-on galaxy with a sharp truncation at a distance of twice the scalelength. Despite the line-of-sight enhancement of the surface bright- 2 DATA ness level in edge-on galaxies, star counts in nearby galaxies are All optical images used in this paper were taken as part of a PhD able to provide significantly deeper profiles with much higher S/N. study on the radial structure of galactic stellar disks by Pohlen Using this completely independent method, Ferguson et al. (2006) (2001). Originally, the disk galaxies ( 2 6 T 6 7) were cho- − followed the profile of the face-on galaxy M 33 down to µlim 30 sen according to the allocated observing time, the observatory ′′ ∼ ′′ I-mag/✷ , which is currently unreachable with surface photom- (north/south), and the available FOV (D25 >2 ) to meet the fol- etry. Their result shows that the broken exponential description is lowing morphological selection criteria verified∼ by using images valid down to the last bin above the noise level and an actual edge from the Digitized Sky Survey (DSS): 1) The galaxies are edge- ◦ of the disk is not found, at least for this galaxy. on with i > 86 . 2) Undisturbed with smooth photometric features Nevertheless, there still seem to be some open issues for a (see below).∼ 3) They look similar to some prototypical cases like complete reconciliation of the renewed face-on with the ‘old’ edge- NGC 4565 or IC 2531, this is done to make it possible to consis- on view (de Zeeuw 2007) as to provide a consistent picture of the tently fit a simple disk model.